Ketogenic Diets 2: Preventing Muscle and Bone Loss on Ketogenic Diets

We’re in the midst of a series exploring therapeutic ketogenic diets. Our immediate goal is to help the NBIA kids, Zach and Matthias, but most of the ideas will be transferable to other conditions – and even to healthy people who engage in occasional or intermittent ketogenic dieting for disease prevention.

Clinical ketogenic diets often produce stunted growth and bone and muscle loss. Today I want to look at this phenomenon and what we can do to avoid it.

Bone Failure and Stunted Growth

First, some data. A review of childhood epilepsy patients on ketogenic diets prescribed by Johns Hopkins Hospital doctors points out problems experienced by the children:

  • Weak bones. Skeletal fractures occurred in 6 of 28 children following the ketogenic diet for 6 years; 4 children had fractures at separate locations and times. [1]
  • Stunted growth. By the end of the 6 years, 23 of the 28 children were in the bottom tenth by height of their age group. [1]

Other negative effects highlighted in the review include kidney stones (7 children developed stones) and dyslipidemia (total cholesterol as high as 383 mg/dl). [1] As we’ve discussed in previous posts, these are probably caused by malnutrition. Kidney stones are usually due to deficiency of antioxidants; dyslipidemia due to deficiency of minerals, vitamins, or choline.

It’s a little hard to nail down the exact cause of the bone fractures and stunted growth because the diets were so atrocious.

First, children were told to eat calorically restricted diets to invoke the starvation response:

Calories were restricted to 75% of estimated daily needs, and fluids were calculated at 80% of daily requirements. [1]

Second, some of the children were fed formula – not real food:

[C]hildren fed only with formula all received a combination of Ross Carbohydrate-Free, Mead Johnson Microlipid, and Ross Polycose formulas to provide a nutritionally complete diet … [1]

For those keeping score, Ross Carbohydrate Free consists of soy protein isolate, high oleic safflower oil, soy oil, and coconut oil, plus vitamins and minerals. Microlipid is a safflower oil emulsion. Ross Polycose is hydrolyzed cornstarch.

(As Jake mentioned in the comments, another commonly prescribed formula is Ketocal, which consists of hydrogenated soybean oil, dry whole milk, refined soybean oil, soy lecithin, and corn syrup solids. Jake’s pithy analysis: “You might as well hold a gun to the head of the child and pull the trigger.”)

There are two problems with this diet design. First, purified diets are notoriously unhealthy; they are missing all kinds of helpful compounds found in real food. Animals do poorly on such diets, as Chris Masterjohn recently noted. Chris quotes the American Institute of Nutrition:

Purified diets without added ultratrace elements support growth and reproduction, but investigators have noted that animals exposed to stress, toxins, carcinogens or diet imbalances display more negative effects when fed purified diets than when fed cereal-based diets.

The second problem, from my point of view, is that they made little use of short-chain fats and ketogenic amino acids to make the diet ketogenic. Instead, they relied on protein and carb restriction and overall calorie restriction to force ketone production. In short, they intentionally starved the kids.

Obviously, starvation tends to produce stunted growth; this is why North Koreans are shorter than South Koreans.

I believe such starvation is totally unnecessary. Use of short-chain fats and ketogenic amino acids can trigger high ketone production even on a nourishing diet.

Nevertheless, even an awful diet is better than the best pharmaceutical drugs:

All of the parents interviewed preferred the diet over medications; 12 cited fewer side effects (such as cognitive dulling, sedation, ataxia, and behavioral problems) from medications that were successfully discontinued, and 11 cited decreased seizure frequency over medications as their primary reason. [1]

Muscle Loss

Another, closely related, problem on ketogenic diets is loss of muscle. You don’t often see bodybuilders or Olympic weight lifters who eat a continuously ketogenic diet. It can be hard to add muscle, especially on protein and carb restricted diets.

This is true even if the diet is not calorically restricted. Which brings us to a rat study [2] discussed by CarbSane in her post “Ketogenic Diet increases Fat Mass and Fat:Total Body Mass Ratio”.

The study compared two diets, a control diet and a ketogenic diet:

The ketogenic diet had more than 6 times the fat of the control diet, the same amount of protein, and no carbohydrate at all. Since protein has to be converted to glucose on zero-carb diets, this ketogenic diet is actually protein restricted. The paper confirms that the ketogenic diet operated on the margin of severe protein deficiency:

[P]reliminary experiments using a KD with 20% protein (as used in children) caused undernutrition of the rats as shown by a significant loss of weight and hair (data not shown). For this reason we used 24% protein, equivalent to that used in controls. [2]

The ketogenic diet also had lower micronutrient levels (“Ash” and “Vitamin”) than the control diet, and much higher omega-6 levels.

Rats were fed ad libitum, meaning they could eat as much as they liked; they chose to eat twice as many calories on the ketogenic diet. This suggests that the diet was protein+carbohydrate deficient.

When protein+carbohydrate intake is deficient, muscle will tend to be catabolized for protein. This causes muscle loss. Meanwhile, the starvation response – especially when more calories are eaten – tends to lead to fat mass gain.

Muscle loss and fat mass gain are exactly what happened to these rats. This is from Figure 1:

You can see in panel A that rats on the control diet weighed more than rats on the ketogenic diet. But panel B shows that rats on the ketogenic diet had more white adipose tissue (WAT). The ketogenic diet rats had more fat mass but less body mass; they had obviously lost muscle and bone mass.

I believe this is due to eating too little protein and carbohydrate. Protein+carbs were 13% on the ketogenic diet, 75% on the control diet. 13% is just too little. For humans, we recommend a minimum protein+carb intake of 600 calories per day, which is about 30% of calories for a sedentary adult. Rats kept in shoebox cages are, of course, sedentary whether they would like to be or not.

I draw two conclusions:

1.      If you’re deficient in protein+carb, you’ll lose muscle; and

2.      Losing muscle may invoke the starvation response, causing you to gain fat.

The paper did not measure length of the rats, but I would bet that the ketogenic diet rats were not only lighter, but shorter as well.

Like the children on Johns Hopkins Hospital’s diet for epilepsy, these malnourished rats experienced stunted growth.

What is the alternative?

As we discussed in the first post in this series, Ketogenic Diets, I: Ways to Make a Diet Ketogenic, there are 3 ways to make a diet ketogenic. One of them is severe protein+carb restriction, but the other two – short-chain fat consumption and supplementation of the ketogenic amino acids lysine and leucine – can generate ketosis even if substantial carbs and protein are eaten.

So it’s worth exploring: with consumption of these ketogenic nutrients, plus substantial carbs and protein, can the health impairments of clinical ketogenic diets be avoided?

Via Nigel Kinbrum comes an interesting paper [3] exploring the use of branched-chain amino acids as an adjunct to ketogenic diets for epileptic children. Most branched-chain amino acids are ketogenic, so this is a good test of my hypothesis.

The study supplemented 45.5 g leucine, 30 g isoleucine, and 24.5 g valine to 17 epileptic children on the ketogenic diet. Leucine is ketogenic, valine glucogenic, isoleucine can be either. The results:

None of our patients had a remarkable reduction in the level of urine ketosis after the supplementation of branched chain amino acids. Moreover, no exacerbation of seizures in terms of frequency or intensity was noted in any of the 17 patients of the study.

Regarding the improvement of seizures, we found 3 patients who had already achieved a reduction of seizures on the ketogenic diet to experience a complete cessation of seizures, while 2 other patients had a further reduction of seizures from 70% on ketogenic diet to 90%. In 2 other patients, the percentage of improvement with the branched chain amino acids supplementation was even greater, achieving 50% and 60% before branched chain amino acids supplementation to 80% and 90% afterward. One patient had 50% improvement (Table 1)….

According to the parents’ and teachers’ reports, improvement was noted regarding behavior and cognitive functions in 9 of 17 patients, particularly in the fields of concentration, learning ability, and communication skills with other children. It is remarkable that 1 of our children had improved so much that she is now applying to attend normal grade level for her age. [3]

There were no significant side effects; only a transient elevation of heart rate at the start of supplementation.

Importantly, supplementing these amino acids allowed more protein to be consumed for the same degree of ketosis:

The first observation we made was that by adding the branched chain amino acids, the fat-to-protein ratio of the diet changed from 4:1 to around 2.5:1 (depending on the patient’s weight) without causing any alteration in ketosis. [3]

Toxicity of Ketogenic Amino Acids

It may be possible to go higher than 45 g leucine per day. The authors acknowledge that they were being cautious in limiting branched-chain amino acid supplementation to that dose:

There is also the question of why we did not try to further increase the amount of branched chain amino acids supplementation because there were no side effects or a change in ketosis. As far as we know, it is the first time branched chain amino acids have been used in patients with epilepsy and we had to be very cautious with their administration. [3]

There is a risk of toxicity at high doses of leucine supplementation unless it is accompanied by the other branched-chain amino acids, isoleucine and valine:

Could we provide leucine alone as the most ketotic of branched chain amino acids? Providing exclusively leucine as an adjunctive treatment to ketogenic diet is impossible because it is toxic when consumed out of proportion to valine and isoleucine…. Lack of valine and isoleucine inhibits protein synthesis. The consequence is that leucine should not be consumed in large amounts without valine and isoleucine, even though only leucine promotes protein synthesis. [3]

Possibly this assessment is over-pessimistic: in rats leucine and isoleucine without valine had no significant toxicity at 5% of energy. [4] Leucine alone lacked toxicity in rat studies:

Recent studies in rats demonstrate no obvious toxicity, even with the administration of BCAA in doses that greatly exceed probable human intake. [5]

L-leucine, administered orally during organogenesis at doses up to 1000 mg/kg body weight, did not affect the outcome of pregnancy and did not cause fetotoxicity in rats. [6]

Lysine, the other purely ketogenic amino acid, is generally considered to have no significant toxicity. [7]

Considerations for the NBIA Kids

For the NBIA kids, Zach and Matthias, we want the diet to be as ketogenic as possible. This is important because glucose is unable to feed neurons due to the inability to make CoA in mitochondria and bring pyruvate into the citric acid cycle. If only ketones can feed the brain, it’s important to make as many of them as possible.

So we would like to give a lot of lysine and leucine. If we have to add other branched-chain amino acids to avoid leucine toxicity, it would be better to add isoleucine, which can be ketogenic, than valine which is only glucogenic.

The BCAA-for-epileptic-children paper [3] can help us judge safe dosages. Supplemental leucine can be at least 45 g/day, since that was give successfully to the epileptic kids. Lysine can be at least as much, since it is non-toxic. Already we’re up to around 400 calories from supplemental lysine and leucine, which is a healthy amount.

Is it necessary to give a lot of isoleucine and valine with leucine? That’s unclear. Leucine by itself may have special benefits for NBIA/PKAN kids.

Paper [5] shows an interesting set of reactions in the brain:  leucine plus pyruvate can be transformed into alpha-ketoisocaproate plus alanine in brain mitochondria. This is extremely important, perhaps, because removing pyruvate from brain mitochondria might prevent iron accumulation in the brain.

Iron accumulation in PKAN is thought to result from pyruvate buildup in mitochondria. Pyruvate attracts cysteine, because pyruvate and cysteine are normally converted to downstream products with the aid of the PanK2 enzyme that is lost in PKAN. With the loss of PanK2, pyruvate and cysteine build up, and the cysteine chelates iron, trapping it in brain mitochondria.

If leucine can remove pyruvate from brain mitochondria, it may also diminish cysteine levels and therefore reduce iron trapping in mitochondria. The iron buildup that is so debilitating might be prevented or mitigated.


I believe the extreme limits on carb and protein intake in conventional clinical ketogenic diets are responsible for their growth stunting, muscle destroying, fattening effects.

In order to supply sufficient protein and carbs while maintaining ketosis, it is necessary to provide ketogenic short-chain fats and amino acids.

Clinical testing of such supplemented diets has so far produced encouraging results. Providing supplemental amino acids to epileptic children on ketogenic diets improved their health and allowed them to maintain ketosis with higher protein intake. Seizure frequency was reduced even as side effects diminished.

Personally, I wouldn’t attempt a long-term ketogenic diet without the aid of coconut oil (or MCTs), lysine, and the branched chain amino acids.

For the NBIA/PKAN kids, it seems that the amino acid supplements should be some mix of lysine, leucine, isoleucine, and valine, with the isoleucine and valine included solely to reduce leucine toxicity. The optimal amount of isoleucine and valine should be smaller than is found in branched-chain amino acid supplements, since leucine by itself may help prevent iron accumulation and increase ketosis. Also, one rat study [4] indicates that isoleucine alone, excluding valine, might be enough to relieve leucine toxicity. Excluding valine would increase the ketogenicity of the supplement mix.

I think the NBIA/PKAN kids will need to experiment with primarily lysine and leucine, and secondarily isoleucine and BCAA supplements, to see what mix works best for them.


[1] Groesbeck DK et al. Long-term use of the ketogenic diet in the treatment of epilepsy. Dev Med Child Neurol. 2006 Dec;48(12):978-81. Hat tip CarbSane.

[2] Ribeiro LC et al. Ketogenic diet-fed rats have increased fat mass and phosphoenolpyruvate carboxykinase activity. Mol Nutr Food Res. 2008 Nov;52(11):1365-71. Hat tip CarbSane.

[3] Evangeliou A et al. Branched chain amino acids as adjunctive therapy to ketogenic diet in epilepsy: pilot study and hypothesis. J Child Neurol. 2009 Oct;24(10):1268-72. Hat tip Nigel Kinbrum.

[4] Tsubuku S et al. Thirteen-week oral toxicity study of branched-chain amino acids in rats. Int J Toxicol. 2004 Mar-Apr;23(2):119-26.

[5] Yudkoff M et al. Brain amino acid requirements and toxicity: the example of leucine. J Nutr. 2005 Jun;135(6 Suppl):1531S-8S.

[6] Mawatari K et al. Prolonged oral treatment with an essential amino acid L-leucine does not affect female reproductive function and embryo-fetal development in rats. Food Chem Toxicol. 2004 Sep;42(9):1505-11.

[7] Tsubuku S et al. Thirteen-week oral toxicity study of L-lysine hydrochloride in rats. Int J Toxicol. 2004 Mar-Apr;23(2):113-8.

Leave a comment ?


  1. I am sure this is a very welcome post for some parents. Would consuming BCAAs disrupt autophagy?

  2. Hi Jeremy,

    It does reduce autophagy temporarily. E.g.

    I think for most of us healthy persons intermittent protein is best — some periods of high protein /BCAA / leucine intake for muscle building, separated by periods of low protein for autophagy / immunity / longevity.

    Best, Paul

  3. Thanks Paul!

  4. Hi Paul,

    the usual commercial BCAA-ratio is 2:1:1 for Leucine:Isoleucine:Valine, so with my 30g/day I have 15g:7.5g:7.5g.
    What ratio would you recommend?
    And what about the Lysine? How much?
    Does health-status matter?

  5. Hi Paul, Great post again! I have managed to order all the supplements and now the leucine and lysine (not easy to do in Denmark). So I should have them next week.

    My question is regarding the diet ratios. I have been adding up to 4 tablespoons of coconut oil per day into Mathias diet. (all recipes and suggestions for this are very welcome! we are going to try the puffed rice coconut bark when my order gets here) I have also been adding in more starch in the form of white rice or white potatoes. However, I still end up with something like Fat: 77%, Protein: 15%, Carbs: 7% with calories averaging around 1500. (He is eating less calories as time goes on.) These are an average over the last few weeks.

    Given your comments today and with the addition of the amino acids and coconut oil, can you give a suggestion of what a 2.5:1 diet looks like? Is that what we are aiming for? I measure ketones everyday so it will be easy to see when/if he falls out of ketosis.

    In all of my research about this disease, I have been struck with how similiar the brains of people with the various neurodegenerative diseases and aged people look. (lots of iron build up and/or damage in various locations in the brain) It seems like if we follow the SAD that we may all end up with these types of problems eventually. When you made the comment “and even to healthy people who engage in occasional or intermittent ketogenic dieting for disease prevention”, it reminded me that this diet is not just for people who are sick but also for people who would like to avoid being sick. Alzheimer’s, Parkinson’s, dementia, etc. are not the fates that any of us would look forward to in our future.

    Thanks again for a great post!


  6. Hi Paul,

    Im a little confused now what is the best way to gain muscle mass. In your book you say only a little protein is needed and overall calories are most important no matter where you get them. Now you’re saying you need enough carbs+protein or high fat intake is only going to make you fat.
    Is the best way to eat at least 600 calories from carbs+protein and then a lot of fat?

    If you eat under 600 calories carbs+protein can coconut oil compensate the deficit? Thanks!

  7. anand srivastava

    From my reading of PHD, I hope Paul agrees. I have come to the conclusion that 1:1:4 is the ideal breakup of protein carbs and fat. With carbs + protein having a minimum level of 150gms.
    For fat loss, reduce fat by an amount that feels effortless.
    For any glucose intolerance keep carbs from 50gms to 100gms.
    This should work for muscle gain also, as you want to eat more for weight increase, providing a corresponding protein increase.

  8. Great post, Paul.

    Since meats and dairy are the best healthy source of branch chain amino acids, it makes sense that a modified Atkins diet for epileptics works wonders too (although likely not as ketogenic as what you propose…and not as tasty as the Spanish Ketogenic Mediterranean Diet (Nutr J. 2008 Oct 26;7:30)).

    My only concern in supplementing so much protein calories in the form of BCAA supplements is the concern you’ve raised recently about micronutrient deficiencies. The same amount of protein from real food would contain vitamins and minerals (although less ketogenic). Maybe you can add a comment about how one should deal with this if you haven’t already (I guess through a higher dose multiple vitamin, but it’s your blog, you suggestion to make. Or not).

  9. Hi Kindy,

    I’m not convinced a 2.5:1 diet is optimal. Remember those are in grams and carbs are zero, so that’s close to 80% fat 20% protein which I think is still too low in protein+carbs.

    Your ratio of 77:15:7 is very good and would be fine if he was eating more calories, but on only 1500 calories a day is a little low in protein and carbs. Hopefully when he feels better he will be more active and eat more. For now, I think we want to experiment to see if we can get him more protein and still keep him in ketosis.

    Ketosis is not an on/off state but a continuous scale of more or less ketone production. We don’t know how strongly ketogenic the diet needs to be to optimize therapeutic effects. You may want to satisfy all the brain’s needs with ketones if you can, which requires a LOT of ketones. So we have to work hard to maximize ketones while not creating nutrient deficiencies, which as Poisonguy points out is a real concern.

    This will require trial and error, I don’t think there is a formula out there yet for the best NBIA diet. I would start from the basic Perfect Health Diet ketogenic diet formula — 200 carb calories, 400 protein calories, plus as many ketogenic short chain fats and amino acids as you can give, plus a nourishing diet including supplements. Then we can vary from there, maybe toward lower carbs and more protein.

    Best, Paul

  10. Hi Franco,

    I guess I have to ask back: What are you eating this stuff for?

    If you want maximum muscle, I would probably add a bit of leucine to your BCAA mix and call it a day. Then eat well.

    If you want a ketogenic diet, I would ask what it is therapeutic for. There are a lot of trade-offs in ketogenic diet design, so it’s not easy to make a one size fits all prescription.

    Hi JVN,

    You need at least 600 carb+protein calories to maintain muscle, since your body will use 600 glucose calories a day and it comes from dietary carbs and protein, or from cannibalizing muscle protein.

    But since an extra pound per week of muscle contains only a small amount of protein, you can add muscle with only a little over 600 carb+protein calories if you give the right stimulus.

    In this post, the rats are deficient in carb+protein. So they’re down well below the 600 calories a day, and more protein or carbs would really help them.

    For people above the 600 calories a day threshold, there are no longer physical shortages of protein that prevent muscle synthesis, but there are hormonal considerations. Extra protein, or specifically leucine and BCAAs, will stimulate muscle synthesis.

    For optimal health, I would do this intermittently, as my reply to Jeremy suggests. I don’t think it’s optimal to be permanently rich in protein or ketogenic.

    Of course the NBIA kids have no choice but to be permanently ketogenic – they can’t metabolize glucose so if their brains are going to work, it has to be on ketones.

    For healthy people I don’t see any reason to risk going under 600 carb+protein calories even if lots of coconut oil is eaten. For the NBIA kids, it may be worth experimenting with, but Kindy will have to be the judge of whether gains outweigh losses.

  11. Hi anand,

    1:1:4 is a good ratio if it’s in calories. In grams the optimal proportion of fat is much lower.

    Yes, you have our view right!

    Hi Poisonguy,

    Yes, great point, micronutrition is very important and as this post points out, purified diets aren’t as healthy as real food diets.

    So we have a trade-off: micronutrition from real food versus the high ketogenicity of purified amino acids. This is an easy choice for nearly everyone – take the real food! – but for the NBIA kids who are genetically incapable of metabolizing glucose, ketones are a brain-saver and high ketogenicity becomes the over-riding concern.

    Also, for athletes who burn a lot of calories, you can eat substantial amounts of real food and still supplement with leucine/BCAAs without harming micronutrition.

    So I would limit the amino acid supplements to these cases, or to some temporary therapeutic “ketogenic diet days.”

    Best, Paul

  12. Franco,

    To prevent having to add Leucine to your BCAA, you could use ModernBCAA by UPSlabs which has a 8:1:1 ratio for Leucine:Isoleucine:Valine.

  13. Paul, I have tried the ‘Mathias’ diet myself and it was difficult for me to eat more than 1200-1500 calories a day. I’m guessing due to the fat. So I’m wondering if I should cycle days? (for instance, one day with higher protein and carbs and far less fat – and the next with higher fat, lower protein/carbs)

    I cannot see getting him to eat more protein and carbs in the current situation (he’s getting about 250 protein and 120 carb calories). The more coconut oil the less he wants to eat. Catch 22. Hopefully the amino acids help this situation at least in regards to the protein.

    I’m guessing from your comments that you believe that the same amount of protein and carbs that is necessary for adults is required by children?


  14. Hi Kindy,

    Distastefulness will lower calorie consumption for a little while, longer if you are overweight, but eventually your desire for calories will rise to match your needs.

    Protein needs scale with body size – they are often quoted as g/kg. Carb needs are hard to assess for Mathias because the biggest glucose consumer, the brain, can’t use glucose in his case. That may lop 300 calories per day from his carb needs. Of course he has to make it up in ketones.

    So, his current intake may be adequate, and maybe that’s why he’s not hungry.

    I think you can trust his appetite for the most part – I wouldn’t try to stuff him against his will. But now that we know what he’s eating, we have to think about how to give him sufficient micronutrition while maintaining high ketosis.

    You might look back at some of our soup recipes in the “Food” category, those would likely be good for him. The Ox Feet broth is a good, fatty soup base and seaweed soups with some eggs and vegetables would be good for him.

  15. Advocatus Avocado


    Thanks for your great blog and book. I’ve purchased the latter for myself as well as a brother.

    I’d like to follow up, if I may, with some of the questions above regarding muscle gain. Specifically, you note that the marginal pound of muscle requires only a modest amount of ingested protein in order to be synthesized. This seems obviously true, but could you speak to what happens to the rest of the muscle tissue when it undergoes the stress of resistance exercise?

    If I perform a heavy deadlift I am damaging an awful lot of muscle tissue throughout my body; the recovery/adaptation to such stimulus is of course what makes me stronger. Does the repair of the original muscle tissue (not the marginal added pound synthesized as part of the adaptation) require additional protein than would otherwise be the case? Or are the components of the muscle tissue entirely recycled?

    I’ve done strength training for a couple of years now, and I’ve noticed that I tend to recover from workouts and increase my strength more quickly when I am consuming more protein (my levels always exceed the minimum threshold you note above). The same seems to be true, anecdotally, among my friends. I am thus somewhat skeptical, despite my sincere respect for your keen analysis.

    Incidentally, I see two references in footnote 11 on p. 25. The first applies to novice lifters, who I suspect on some levels may be qualitatively different metabolically than more advanced lifters. (I’ve seen out of shape people add muscle mass initially even if binging on cheetos.) The second (Hoffman JR et al) seems more applicable to my situation. They of course find that those who consume the greatest proportion of protein have the greatest strength gains, however it’s not statistically significant. (Perhaps the low sample number and 3 month duration are not enough to tease out a difference?) Also of interest in this study is that all 3 cohorts have roughly similar caloric intakes. The other macronutrients are thus not controlled for– indeed, there is a statistically significant difference in carbohydrate consumption across groups. An interesting contrast for this study could therefore be to test the hypothesis you have in the book: namely, comparing diets that are equal up to a certain caloric level, whereupon a set number of additional calories from a given macronutrient are layered in.

  16. Hi Advocatus,


    Protein is recycled effectively, so when muscle is broken down during resistance exercise the amino acids can be re-used to build new muscle. However, the amino acids may not be in the right place at the right time, so having a protein surplus does help.

    In the book we do recommend that athletes and bodybuilders eat at the high end of our plateau ranges – more like 1000 calories carb+protein than 600 — in order to assure the presence of sufficient protein so there are no bottlenecks. Also, this naturally provides more leucine / BCAAs so you get the hormonal / signaling benefits too.

    But I don’t think it’s necessary to go above 1000 calories + maybe 100 calories per hour of intense training. I doubt there are many benefits in terms of strength gain above that level.

    The studies are interesting but I give more weight to experiences like yours and other athletes, because the studies are underpowered. The studies show that effects aren’t large; but of course small gains build up over years of training to large differences — and the separation between the top athlete and the 2nd best athlete is tiny. So athletes can learn more from the experiences of other athletes than from the literature at this point.

    As far as I know, athlete experiences are consistent with our prescription. I’m not aware that going well above our 1000-calorie carb+protein suggestion has major strength benefits. I think there are significant benefits going up to that level.

    I do think studies of athletic performance and diet are very interesting and we should have more of them, and larger more careful studies. Most of the studies I see are really small.

    Best, Paul

  17. CorruptionWatch

    To feed one group of animals a purified experimental diet and call a group on a food-based diet a control is absurd, inexcusable, and constitutes scientific malpractice.

    Papers should be withdrawn. Journals should be questioned. Funding should be redirected to people doing actual scientific research.

  18. Hi CorruptionWatch,

    What about the purified experimental diet fed to the human epileptic children?

  19. Advocatus Avocado

    Hi Paul,

    Thanks very much for your prompt and thoughtful reply. It’s part of what makes this website unique.

    You present a strong case, as always. I’ll just note that I believe my performance improved (albeit marginally–the differences aren’t large) when I allowed my protein/carb/fat ratios to remain consistent despite my high caloric intake, which is ~3,600 calories/day. In other words, I had a sense of better performance when I lowered my fat% to around 65 and allowed around 200g/day of protein (I work out 2-3x a week for an hour). This seems to be a fairly insignificant deviation from your general recommendations, however, so I’ll stop splitting hairs.

    I’d like to reiterate that I’m a great proponent of the PH diet. I’m pleased to report that after dramatically increasing my saturated fat intake and following some of the proscriptive measures in the book that my blood pressure has decreased from 130/85 to 118/74. My resting heart rate has also decreased, and I now sleep much more soundly.

  20. Hi Advocatus,

    That’s great to hear! I’m glad you’re doing well.

    Best, Paul

  21. (As Jake mentioned in the comments, another commonly prescribed formula is Ketocal, which consists of hydrogenated soybean oil, dry whole milk, refined soybean oil, soy lecithin, and corn syrup solids. Jake’s pithy analysis: “You might as well hold a gun to the head of the child and pull the trigger.”)

    I have quoted Jake extensively since he said these wise words. The problem is that tube fed kids get formula paid for by insurance. Real food? Not a chance they would pay for that. So insurance is happy to finance the experiment!


  22. Hi Paul:
    Regarding your recommendation for 4tbs per day of fat(good fats-coconut oil, butter, olive, if you eat avocado on a daily basis( 1/2 avocado for ex) would you still want the 4tbs added to diet on a daily basis?

  23. Hi Steve,

    4 tbsp is just a suggestion, not a hard number. For weight loss diets the first thing we suggest removing are those fats. On the other hand, active people burning a lot of calories can eat as much of those as they want.

    It just happens to work out that for a typical person eating typical foods (i.e. not foods like brain or bone marrow that are very fat-rich and were prized in the Paleolithic), adding 4 tbsp fat makes the macronutrient ratios come out close to what we think is perfect.

    So the spirit behind the 4 tbsp suggestion is: Don’t be afraid of fats; eat them to your taste. If you like butter on your potato, don’t be afraid to slather it on.

    Re avocado specifically, it is a fatty food and an alternative to butter or oil. Guacamole counts like sour cream if you’re counting calories. But I wouldn’t be too concerned about whether you need to reduce other oils because you’ve added an avocado. Eat to taste and your taste buds will guide you properly. They evolved to like fat for a reason.

    Best, Paul

  24. As I pointed out in the other post, Paul, even Cordain is coming around on fats. So lather up, everyone. 😉

  25. Paul,

    my IF (20h fast/4h fed) is ketogenic by default during at least the last 6 hours of the fast, my fruity breath is giving it away.
    The meals in the feeding window are along the standard PHD with a few extra carbs to account for refilling glycogen stores.
    So, not therapeutic for anything special and just would like to gain max muscle with min fat, besides I just like to eat once per day, always did.

    Brian R.,

    thanks for the tip but I’m in europe, even east-europe most of the time, and see no chance to get a regular supply of such rare supps.
    I’m happy already to found a good source of affordable quality bcaa without added anything.
    I had extra leucine once (shipped from the US with all tax/customs problems you could imagine!) but I didn’t feel anything different or saw any different results. Plus it was f**** expensive!

  26. Hey Paul

    I know this is an old post, but it seems like I initially missed it. Either way, in one of the comments you recommend athletes consume 1000 calories of carb+protein making the total intake of carb+protein 250 grams. Now if I was to consume 4000 calories daily I would have to consume the remaining 3000 calories from fat alone? Is this realistic, or do you recommend I stick with a set ratio of 65% fat, 15% protein, and 20% carb, which would ultimately mean I would consume 600 calories from protein alone and 800 from carbohydrates with the rest coming from fat? Which calorie breakdown do you recommend I follow?

    I have been an underweight athlete for years now, and I am constantly under performing and asked by my trainers to put on weight, but I think too much about food, and the effects it might have on my health. Just today I read on a forum that saturated fat above 50% percent could stiffen cell membranes? Is there any truth to that. The constant reading I think is driving me crazy and is preventing me from performing at my optimal level.

    Thanks a lot Paul.

  27. Hi Robert,

    I’ll post on that shortly.

    If you’re really consuming 4000 calories daily you have some leeway; you’ll be running down muscle glycogen fairly significantly each day so you’ll be able to eat more carbs.

    A reasonable caloric guideline for athletes is to eat at the upper end of our carb and protein ranges. That would be 600 carbs + 100 carbs per hour of training + 600 protein calories + fat for the rest. Say you train 2 hours a day, that would be 800 carb calories, 600 protein, and 2600 fat calories. Which is 65% fat, 15% protein, 20% carb — what do you know!

    Don’t worry about saturated fat. Dietary saturated fat does not stiffen membranes. If saturated fat could become a very high percentage of membranes, it would stiffen them, but the body regulates membrane composition and it doesn’t happen. Eat as much saturated fat as you like, your membranes will still be 40% saturated.

    Best, Paul

  28. The 4tbs suggestion is it a US tablespoon which is 15mL, in Australia 1 tbs is 20mL?

  29. Hi Sue,

    Yes, that was a US tablespoon. Again, this is just a “typical person’s diet,” you should eat as much as you like, less or more to fit your needs and tastes. There’s no health benefit to eating more or less fat.

  30. Mct oil versus Coconut oil, Pros, Cons….?
    I consume both MCT oil & coconut oil. Just wondering if you have any views if one is better than the other, or maybe they have different benefits…..(or negatives)?

    Further info;
    I would guess that the coconut oil is less processed than the MCT oil, just a guess though.
    Two of the things i like about the MCT oil are that it is tasteless & remains liquid in colder temps.

    Fat breakdown of each, according to the labels;
    MCT oil: (Saturated 100%) C6 0.5%, C8 65%-75%, C10 25%-35%, C12 <1%

    Extra Virgin Coconut Oil: C8 8%, C10 9%, C12 49%, C14 16%, C16 8%, C18 2%. Saturated 92%. Unsaturated 6%. Polyunsaturated 2%.

  31. Hi Darrin,

    You already broke it down pretty well!

    MCTs are more ketogenic, partly because of the shorter chain length (C8 vs C12) and because coconut oil has long-chain fats too.

    You can think of coconut oil as the equivalent of 50% MCTs, 50% butter.

    So it comes down to taste, cost, and desired ketogenicity.

    Best, Paul

  32. MCT sounds like something I might tolerate. Try as I might, I cannot tolerate the taste or smell of coconut in any form.

    How might I add MCT to the PHD?

  33. Hi erp,

    Taking a few tablespoons when fasting would be a good method.

  34. Can I reach a deep state of ketosis if I don’t use amino acids but only coconut oil and milk combined with a high fat diet (75% fat, 25% protein/carbs)?

    If not, is it enough to just take the branched amino acids in a 2:1:1 percentage without taking lysine?

    My motivation is to heal my migraines.

  35. Hi Primordial49,

    For most conditions it’s not necessary to reach deep ketosis, and it’s usually best to go as little into ketosis as you need to mitigate your problems.

    I actually think it’s best to avoid the lysine entirely and take leucine or BCAAs, 5 g/day, with coconut oil or MCTs.

    Best, Paul

  36. Paul, I’ve been taking L-Lysine or years. Do you think I should switch to leucine or L-leucine? I started taking it for sores in the mouth and tongue. They stopped almost immediately when I started taking it.

  37. Hi erp,

    Yes, I think you should stop it for a while and see what happens. If the sores come back then re-start.

    The leucine is an independent issue. It’s been shown to extend lifespan and help build muscle mass, so I think it’s beneficial. Not very tasty however.

  38. Thanks for your answer, Paul!

    There are books like that out there (Primal Body, Primal Mind, for example) which promote ketosis as a beneficial state to live on for the entire life. Did you have writtena about your stance on why ketogenic diets aren’t beneficial in the long-term if prepared properly?

    Otherwise it would be great to hear why it is best to go as little into ketosis as needed. From my perspective ketosis is one of the greatest things available to us diet wyze.

  39. Hi Primordial49,

    I discussed this briefly in my AHS review of Nora Gedgaudas’s talk. While ketones are great for neurons and are one of my favorite therapies for neurological disorders, they do promote growth of eukaryotic pathogens and so you if you generate an excess of ketones that is getting excreted, it’s doing little good to the neurons but potentially helping pathogens grow into a serious infection.

  40. Hi Paul/Others;
    I have a Leucine related question;

    I did a quick google for

    & received seemingly contrary answers;
    ie. ‘leucine increases insulin sensitivity’ or ‘leucine reduces insulin sensitivity’.
    also, ‘leucine decrease insulin resistance’ or ‘leucine increase insulin resistance’.

    Anyone know which may be nearer the truth in humans?
    Or may be both are correct….. dependant on the leucine dose &/or usage (chronic leucine supplementation).

  41. Hi Darrin,

    It looks to me after a brief scan of the literature that leucine induces insulin resistance in the liver but insulin sensitivity elsewhere.



  42. Thanks Paul,

    I was looking at Leucine to see if it could be used in some way to help me with fat loss; getting rid of those last 1 or 2 stubborn kilos.
    But, i am seeing mixed messages about Leucine helping with fat loss. Plus when & how to use Leucine to help with fat loss.
    So i think i’ll give the leucine a miss for now as i cannot be sure if it will help, hinder, or neither, my fat loss (plus its not that cheap).

    Also, i’ve just found your post titled ‘Perfect Health Diet: Weight Loss Version’ here
    So i’ll have a read through that for now & see what tweaks i can implement & keep it simple.

  43. Hi Paul,
    We started the Gaps diet with my 4 year old last helped her mild autistic symptoms, but resulted in extreme muscle wasting and weight loss. We added rice and she improved a bit, but still had gut issues..
    We are certain she has Hasimotos due to gut issues ..she tires uickly, hair falling out,…Do have any strater ideas for the keto diet and can you give proper supplement dosages for the amino acids you mention?
    She doesnt tolerate alot of foods..due to leaky gut..
    thanks KATE

  44. Hi Kate,

    First, she does need glucose. You should experiment with dextrose powder, rice syrup, and tapioca syrup to see if they are better tolerated than rice.

    I would not necessarily recommend a ketogenic diet for Hashimoto’s; it might work for some infection types but certainly not all. I think a 30% carb diet would be safer.

    So I wouldn’t supplement amino acids like leucine. Rather just eat whole foods.

    Bone broth soups and a low omega-6 diet, with an emphasis on fish, shellfish, and beef/lamb, would be a good start, plus sufficient carbs.

    Best, Paul

  45. Karen McKenzie

    Hi Paul
    I have been interested in the ketogenic diet for years as have a daughter (spina bifida/hydrocephalus) with epilepsy which although has responded to medications, she still has grand mal seizures once every 2-3 weeks (at least she is no longer hospitalised when she has these unlike in her pre-school years). I am not happy with the side effects however, as the lamotragene has caused dizzyness and vomitting. She has been on a gluten free diet for the past year as has been tested and is gluten intolerant. She has also always had an addiction to carbs (mainly bread), although she doesn’t eat sweets, drinks only milk/water. I’m afraid with the gluten free diet all I have switched is wheat for rice flour, etc, breads so she is still getting quite high carbs. I’ve looked at the Modified Atkins Diet for her, but after buying your book am very interested in your take on the ketogenic diet (we use a lot of coconut oil already)and using amino acids so the carb intake can be upped a bit. There are so many really low carb bread recipes using coconut flour, etc, on the web now that I feel I may be able to tackle a ketogenic diet without her rebelling. However, I just want to clarify and make sure I am on the right track with the amino acids before I order them – 45g lysine, 45g leucine, 30g isoleucine, 24.5 g valine as mentioned in the study on epileptic children. Should she use MCT oil rather than coconut oil to up her ketosis. She is now a young adult and I feel that I have procrastinated long enough about starting this diet, so any other suggestions would be greatly appreciated.
    Thank you.

  46. Hi Karen,

    Sorry, I didn’t make it clear in the post, but I don’t think one should take so much amino acids.

    In fact, I wouldn’t recommend taking lysine at all, and I think 5-10 g BCAA containing 50%-80% leucine, 10%-25% each isoleucine and valine, would be perfect along with however much MCT oil was needed to reduce seizures. You have to experiment a bit to see if it works and how much MCT oil you need.

    You’ll need less MCT oil than coconut oil, maybe half as much, but if she prefers coconut oil then use that. MCT oil is pretty tasteless and inexpensive.

    I think the key thing is to have her eat whole foods – not commercial prepared foods. If she can learn to cook and adapts her taste to fatty non-sugary foods, then it will stand her in good stead for the rest of her life.

    Best, Paul

  47. Karen McKenzie

    Paul, Thanks for your prompt reply. I was a bit confused about how many amino acids she should have and worried about this part of the protocol, hence the post to you, so thank you very much for the clarification. I do agree wholefoods are better and she enjoys helping me cook (but definitely more on the baking side!) If she can still have noodles in the form of rice noodles and perhaps spring rolls with rice sheets and homemade low carb muffins/bread, I think she will handle the rest of it. I have a week’s leave next week so am hoping to be organised and start this.
    Once again, thanks for getting back to me, it never ceases to amaze me how busy people like yourself go out of the way to answer questions.
    Kind regards

Leave a Comment

NOTE - You can use these HTML tags and attributes:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

This site uses Akismet to reduce spam. Learn how your comment data is processed.